AI research is becoming difficult to separate from geopolitics

Machine Learning


top of the world The AI ​​research conference, the Conference on Neural Information Processing Systems (better known as NeurIPS), this week became the latest organization to become embroiled in the growing conflict between geopolitics and global scientific cooperation. Conference organizers announced controversial new restrictions on international attendees, which they quickly rescinded, after Chinese AI researchers threatened to boycott the event.

“This is a potential tipping point,” said Paul Triolo, a partner at DGA Albright Stonebridge, an advisory firm that studies U.S.-China relations. Although Toriolo has argued that attracting Chinese researchers to NeurIPS is good for U.S. national interests, some U.S. officials have called for U.S. and Chinese scientists to separate their research, especially in the field of AI, which has become a particularly sensitive topic in Washington.

The incident deepens political tensions over AI research and could deter Chinese scientists from working at U.S. universities and tech companies in the future. “At this point, it will be difficult to keep basic AI research out of society at some level. [political] It’s a photo,” says Toriolo.

NeurIPS organizers announced the latest restrictions on participation in their annual handbook for paper submissions, published in mid-March. The rules state that the event cannot provide services, including “peer review, editing, or publication,” to any organization that is subject to U.S. sanctions, and is linked to a database of sanctioned organizations. This included companies and organizations on the Bureau of Industry and Security’s Entity List, as well as companies and organizations on another list with suspected ties to the Chinese military.

The new rules could affect researchers from Chinese companies such as Tencent and Huawei, who regularly publish their research on NeurIPS. This database also includes organizations from other countries such as Russia and Iran. The United States has restrictions on doing business with these organizations, but there are no rules regarding academic publishing or conference participation.

The NeurIPS Handbook has since been updated to specify that this restriction applies only to Specially Designated Nationals and Blocked Persons, a list primarily used by terrorist groups and criminal organizations.

“In creating the NeurIPS 2026 Handbook, we included links to the U.S. government’s sanctions tools that cover a much broader set of restrictions than those that NeurIPS would actually have to follow,” event organizers said in a statement released Friday. “This error was due to miscommunication between the NeurIPS Foundation and our legal team.”

Before reversing tack, conference organizers initially said the new rules “concerned the legal requirements applicable to the NeurIPS Foundation, which is responsible for complying with sanctions,” adding that they were seeking legal consultation on the issue.

immediate backlash

The new rules immediately sparked a backlash from AI researchers around the world, especially from China, which produces a large number of cutting-edge machine learning papers and has a growing share of the world’s top AI talent. Several local academic organizations have issued statements condemning this measure and, more importantly, discouraging Chinese scholars from joining NeurIPS in the future. Some argued that it could encourage Chinese academics to contribute to domestic research councils instead, increasing the country’s influence in related science and technology fields.

The Chinese Association for Science and Technology (CAST), an influential government organization for scientists and engineers, announced Thursday that it would stop funding Chinese scholars who travel to participate in NeurIPS, and instead use the funds to support domestic and international conferences that “respect the rights of Chinese scholars.”

CAST also announced that it will no longer count papers from the 2026 NeurIPS conference as academic achievements when evaluating future research funding. Now that NeurIPS has withdrawn the new rules, it is unclear whether the organization will change course.



Source link